Abstract #0137

MR Estimation of Longitudinal Relaxation Time (T1) in Spoiled Gradient Echo using an Adaptive Neural Network

Hassan Bagher-Ebadian^1,2, Siamak P. Nejad-Davarani^1,3, Ramesh Paudyal¹, Tom Mikkelsen⁴, Quan Jiang^1,2, James R. Ewing^1,2

¹Neurology, Henry Ford Hospital, Detroit, MI, USA; ²Physics, Oakland University, Rochester, MI, USA; ³Biomedical Engineering, University of Michigan, Ann Arbor, MI, USA; ⁴Neurosurgery, Henry Ford Hospital, Detroit, MI, USA

Estimating the longitudinal relaxation time, T1, from spoiled-gradient-recalled-echo (SPGR) images is challenging and susceptible to the level of noise-to-signal ratio (SNR) in acquisition. Methods such as Simplex-Optimization, Weighted-Non-Linear-Least-Squares, Linear-Least-Square, and Intensity-based-Linear-Least-Square have been employed to estimate T1. In linear and non-linear methods, the estimated T1 values are dependent on defining the weighting factors, which may result in a biased estimation. Herein, an adaptive neural network is trained and compared with different techniques using an analytical model of the SPGR signal in the presence of different levels of SNR. Receiver-Operator-Characteristic analysis and K-fold-cross-validation were employed for validation, testing, and network optimization.